On Rank Distributions of Letter Frequencies in Human Languages

The dependence between letter frequencies and letter ranks (positions in the alphabet reordered by decrease of frequencies) in human languages was described both as a linear function of rank and frequency logarithm and as a linear function of rank logarithm and frequency. Both models were used to draw analogies with other information systems. We demonstrate that neither model is superior as compared to the other one, and thus the problem of rank distribution of letter frequencies remains open.